Structural Insights into Functional Impacts of Common Variants in Mitochondrial Protein-Coding Genes of Mediterranean Loggerhead Turtles

The Mediterranean population of the loggerhead turtle (Caretta caretta) originates from a few colonization events from rookeries on Oceanic beaches. Mediterranean loggerhead turtles may have unique genetic adaptations to the region climatic conditions, due to their temperature sensitivity, affecting various biological functions. We used complete mitochondrial DNA sequences from 61 independent individuals sampled in the Mediterranean to infer the protein-coding variants. The 3D structures of the subset of proteins affected by non-synonymous substitutions were reconstructed to hypothesize the ensuing effects for the protein functionality from a structural and energetic point of view. By performing two consecutive sets of comparisons between proteins encoded by Hg IB vs. basal Hg II and Hg IB vs. derived Hg II we gained insights on whether the new variants replicate and potentiate the evolutionary trend observed in the long-term divergence between Hg IB and Hg II. Minor changes in protein stability were predicted in the Hg IB vs Hg II comparison, consistent with the long-term evolutionary viability of the amino acid substitutions in the two lineages. The five comparisons involving new variants, derived within Hg II, predicted a slight destabilization of the corresponding protein structure within the mitochondrial membrane 3D context, while drastic effects on the proteins’ functionality could be ruled out. Our analysis provides a novel view of the evolutionary dynamics of mitochondrial DNA and the potential functional implications of specific mutations associated with the colonization of the Mediterranean, contributing to a deeper understanding of the genetic diversity within and among C. caretta haplogroups.

The loggerhead turtle ( Caretta caretta ) is the most common marine turtle in the Mediterranean Sea, where it is found both as a temporary foraging visitor and as a regularly nesting species (Casale and Margaritoulis 2010). However, its home range is much broader, as it inhabits tropical and subtropical waters of the Atlantic, Pacific and Indian Oceans (Márquez 1990). Current views regard the Mediterranean population(s) as originating from a limited number of colonization events by individuals from rookeries on Oceanic beaches (Bowen et al. 1993; Encalada et al. 1998; Clusa et al. 2013; Baltazar-Soares et al. 2020). This species displays a natal homing behaviour by which adult females mate in open waters and land to lay their eggs on or near the beaches where they were born. A low level of breakdown of strict homing behaviour, e.g. by ”social facilitation” (Bowen et al. 1992; FitzSimmons et al. 1997) or straying (Shamblin et al. 2015) can maintain residual connectivity among rookeries as well as enabling the colonization of new breeding locations and eventually the establishment of new colonies. This process is now considered to occur more often than previously thought (Garofalo et al. 2009; Shamblin et al. 2014; Shamblin et al. 2015; Carreras et al. 2018). Populations breeding on Mediterranean beaches and foraging in Mediterranean waters for thousands of generations could potentially benefit from adaptations to the climatic conditions of this Sea, and thus display the rise in frequency of novel positively selected variants (Somero 2010). During the colonization process, this species must have resisted the intervening climatic fluctuations, an outcome also proposed for some of the other subtropical species currently in the Mediterranean (Wilson and Eigenmann Veraguth 2010). As poikilothermic vertebrates, marine turtles are expected to have been strongly impacted by such fluctuations, given the temperature dependency of many aspects of their biology (phenology, migration, reproduction and sex determination) (Mazaris et al. 2004; Hawkes et al. 2007). In this scenario, the repertoire of protein-coding variants, particularly in mitochondrial genes, may include amino acid substitutions with a specific biological significance. Indeed, multiple aspects suggest that the mitochondrial genome may be deeply involved in the climatic adaptations of this species and could thus represent a valuable target for more in-depth analyses. First, this organelle is central to energy production and heat release. Second, the plasticity of mitochondria and their genomes is involved in promoting adaptation to short- and long-term environmental changes (Breton et al. 2021). Therefore, upon the serial accumulation of variants due to the absence of recombination, this genome may behave as an “adapted genomic region” (Stiebens et al. 2013). Atomic structures and atomic-scale interactions of the membrane and multi-protein complexes in which the mitogenome-encoded products are embedded can be predicted with high confidence from related taxa (Zhu et al. 2016). In this framework, appropriate structural analyses may indicate which amino acid variants found in mitochondrially-encoded proteins might be adaptive in Mediterranean conditions. We recently generated complete mitochondrial DNA sequences from 61 independent loggerhead turtles sampled in the Mediterranean (Tolve et al. 2024). In the present study, the same data have been leveraged to infer the protein-coding variants represented in this dataset after merging them with previously published complete reference sequences (Drosopoulou et al. 2012; Duchene et al. 2012; Hernández-Fernández and Delgado Cano 2018). The 3D structures of the subset of proteins affected by non-synonymous substitutions ( Table 1 ) have been reconstructed to clarify possible alterations introduced by the derived amino acid compared to the ancestral one and to hypothesize the ensuing effects for the protein functionality. This analysis involved, first, the molecular modelling of the mitochondrial subunits containing the amino acids encoded by the so-called Haplogroup (Hg) IB, i.e., the mitochondrial genetic lineage most commonly found in Atlantic loggerhead turtles (Fig. 1). We then replaced these residues with those of the reference FR694649.1, which is a representative of the so-called Hg II, i.e., the mitochondrial genetic lineage of most turtles sampled in the Mediterranean (Duchene et al. 2012). Finally, we replaced additional residues to emulate the subunits encoded by uncommon variants detected in our screening. This sequential approach enabled the generation of new models to analyze the effects of the substitutions from a structural and an energetic point of view. Our working hypothesis was based on the fact that i) the environmental conditions of the Mediterranean Sea can be considered “derived” as compared to the Oceanic home range and ii) Mediterranean breeding colonies are populated by carriers of Hg II, whereas carriers of Hg IB are found as visitors and do not lay eggs on Mediterranean beaches. In this context, a direct comparison between the structures of proteins encoded by Hg IB and II informs on a range of possible adaptive changes that may underlie the preferential successful colonization of the Mediterranean by Hg II-carrying founders. Our previous study (Novelletto et al. 2016), consisting in a limited screening of NADH ubiquinone oxidoreductase subunits ND1 and ND3 variants, revealed amino acid substitutions compatible with homeoviscous adaptation (Hazel 1995; Ernst et al. 2016), a mechanism widely documented in poikilothermic vertebrates (Loftus and Crawford 2013; Strobel et al. 2013). The same study, while highlighting the importance of studying structural variation as an interpretation of experimental data (Falconi et al. 1998; Napoli et al. 2008; Di Marino et al. 2010) prompted(Novelletto et al. 2016) the extension of the same analyses to entire sets of mitochondrial subunits inherited together. Here, we report on additional amino acid substitutions in all subunits encoded by the mitogenome, that define Hg II sublineages and for which the reconstruction of ancestral and derived states is immediate. By performing two consecutive sets of structural comparisons, i.e. Hg IB vs. basal Hg II and Hg IB vs. derived Hg II we gained insights on whether the new variants replicate and potentiate the evolutionary trend observed in the long-term divergence between Hg IB and Hg II.

The sequencing dataset The present analysis is based on the complete mitogenome sequencing set described in Tolve et al. (2024). This consists in mitogenome assemblies from shotgun sequencing of 61 loggerhead turtles sampled in Veneto, Emilia Romagna, Tuscany, Latium, Calabria, the island of Linosa and the strait of Sicily, including dead embryos. Coverage of the mitogenome across specimens ranged 23 to >1,000. Variant Calling Pipeline FastQC 11.9 software (https://www.bioinformatics.babraham.ac.uk/projects/fastqc/) was used to perform quality control on the reads. The primary quality assessment parameter was the base quality score, subsequently employed to calibrate the trimming and filtering. Trimming was performed using TRIMMOMATIC v.0.39 (Bolger et al. 2014) with the following parameters: HEADCROP:10, TRAILING:28, LEADING:28, and MINLEN:15. Adaptors and low-quality reads were removed. All samples passing quality control were subjected to variant calling using a standardized pipeline with consistent parameters. Bowtie2 v.2.4.4 (Langmead and Salzberg 2012) was employed for alignment, referencing the mitochondrial genome of C. caretta (GenBank accession number: FR694649.1). Samtools v.1.13 (Li et al. 2009) and bcftools v.1.13 were utilised to manipulate the alignment file. The Sequence Alignment Map (SAM) file was compressed in Binary Alignment Map (BAM) format, sorted and indexed. Variant calling was performed using bcftools mpileup, and bcftools call commands, incorporating the multiallelic caller option. Variants with low quality were filtered out. Variant annotation was performed with snpEff v.5.2c (Cingolani et al. 2012). The mitochondrial genome reference for C. caretta was constructed using the FR694649.1 annotation. Protein Structure Modeling The molecular modelling of the Atlantic C. caretta subunits harbouring was performed using the Modeller 10.4 software (Webb and Sali, 2016). The SWISS-MODEL web server (Waterhouse et al. 2018) was used to identify homology modelling templates for each subunit. Selection considered sequence identity percentage and the Global Model Quality Estimation (GMQE) score. From 0 to 1, the GMQE score reflects the expected accuracy of a model derived from the evaluated template, normalized by the target sequence coverage (Table 2). Five models were generated for each subunit. The best model was selected based on the Discrete Optimizer Protein Energy (DOPE) score. The program calculates a distance-dependent statistical potential to evaluate the quality of a homology model. Lower DOPE scores indicate superior models (Shen and Sali 2006). To assess the structural impact of mutations, the difference in free energies (∆∆G) between the structures being compared was calculated by FoldX (Schymkowitz et al., 2005). This value reflects the mutation effect on the protein 3D structure from an energetic point of view (Table 3A, B). The FoldX suite predicts the free energy changes between a wild-type (WT) and a mutant (MT) protein, and then calculates the difference: ΔΔG (change) = ΔG (MT) −ΔG (WT) The ΔG(WT) and ΔG(MT) values alone are meaningless, while the calculated ΔΔG(change) has been shown to correlate well with the experimental values (Schymkowitz et al. 2005). The FoldX suite assumes that when ΔΔG(change) is > 0, the mutation is destabilising, while when ΔΔG(change) is < 0, the mutation is stabilizing. The expected margin of error for the FoldX program is approximately 0.5 kcal/mol, so ΔΔG changes in the -0.5 - +0.5 range cannot be considered significant. Analysis of conservation We used blastp as implemented at https://blast.ncbi.nlm.nih.gov/ with protein accessions reported in the FR694649 GenBank file. Search was limited to Chelonidae (taxid: 8465). The retrieved sequences were aligned with COBALT (Papadopoulos and Agarwala 2007), and the resulting alignment inspected visually. Variant notation Throughout the text the notation POSancestral>derived is used for amino acid substitutions that could be polarized, whereas the notation variant1POSvariant2 is used for the remaining substitutions. For DNA, the notation POSancestral>derived is used for nucleotide substitutions that could be polarized, whereas the notation POS variant1/variant2 is used for the remaining substitutions. Phylogenetic position of non-synonymous variants The maximum parsimony tree obtained with the 61 complete sequences of our dataset is displayed in Figure 1 . The dataset included one Hg IB representative, and additional Hg IB and IA reference sequences were used (Tolve et al. 2024). The tree replicated the topology previously described for the species in the context of other Chelonidae (Duchene et al. 2012), with Hg IB as the most basal and Hg IA as a sister clade of Hg II. Within Hg II, 50 variants residing in protein-coding regions were detected, 33 of which were shared by all Hg II carriers and then considered basal to the haplogroup. The 9440A>T synonymous DNA variant (COX3 254Val>Val ) was shared by all but two Hg II carriers, identified as CCA2.9. Based on the tree structure, the ancestral and derived states of the 33 Hg II-specific mutations could be easily polarized by parsimony ( Fig. 1 ). Five non-synonymous substitutions (Table 1) turned out to be basal to all Hg II sequences (red labels in Fig. 1 bottom right). Note that the reference sequence FR694649 turned out to be highly derived and can be considered a representative of the Hg II basal sequence as far as missense variants only are concerned. The total number of missense substitutions differentiating FR694649 from Hg IB amounted to 22 (Table 3A). In the Hg II subtree, seventeen protein-coding variants were polymorphic and defined sub-branches. Among these we detected an event of heteroplasmy, discernible from sequencing errors and consisting in the coexistence of the normal sequence and the truncating variant ND2 234Trp>Stop, an apparently very severe substitution. Based on the strong conservation of this residue and the total protein length in other species, we argue that this observation testifies to an evolutionary dead-end, with no chance for this variant to be transmitted to the offspring in purity. This variant will not be further discussed. Six of the polymorphic variants were non-synonymous ( Table 1 and red labels in Fig. 1 top left) and each recovered in a single specimen, except for the ATP6 135Pro>Ser, found in the two CCA2.9 individuals and the CYTB 53Ala>Thr, found also in two individuals. Therefore, frequency of all non-synonymous polymorphic variants in the Mediterranean population at large is to be considered low, though at least in one case (CCA2.9) the corresponding sub-branch had the opportunity to further radiate, accumulating additional variation (Fig. 1 top left). Among the 11 non-synonymous mutations that are specific to Hg II (Table 1), eight lead to an increase in AT content. This is in accord with a previous comparative analysis of codon usage in different mitochondrial genomes and appears to be a characteristic of the reptile mitogenome (Montaña-Lozano et al. 2023). An analysis of conservation in the homologous positions of other marine turtle species (Table 1) revealed that the five non-synonymous substitutions basal to Hg II hit preferentially positions that are remarkably variable across species, with three instances of multiple amino acid substitutions in different Chelonidae. Conversely, the six substitutions defining Hg II sub-branches hit positions that are more conserved, with three instances of invariant amino acids among Chelonidae. In summary, the above observations are compatible with hypotheses ranging between i) the six derived variants being nearly neutral and currently drifting at low frequencies in the Mediterranean population, and ii) the presence of six derived variants representing a case of accelerated substitution rate(s) in the Mediterranean environment, i.e. a condition resulting from positive Darwinian selection. In order to extract signatures of these alternative scenarios, we modelled the structures of the corresponding proteins with the amino acids observed in Hg IB, as well as the basal and derived amino acid states in Hg II, leading to the considerations reported in the following sections. We reasoned that the ΔΔG values for the comparison between Hg IB and Hg II basal (Table 3A) provide a null for the difference in stability of the proteins encoded by the lineage that shifted from the Ocean to the Mediterranean, whereas the comparison between Hg IB and Hg II including variants (Table 3B ) indicate whether the more recent substitutions replicate or counteract the same trend. Impact of mutations on protein 3D structure ATP6 - ATP synthase Complex V is a large transmembrane protein complex crucial to oxidative phosphorylation. The mt-ATP6 subunit in the mitochondrial inner membrane is an integral component of complex V ( Fig. 2A ) and is essential for protein assembly. Three mutations were identified within the ATP6 subunit. The Ala78Thr and Leu151Phe mutations ( Fig. 2 ) that differentiate Hg IB from Hg II are mapped on α-helices in the transmembrane protein core. The hydrophilic/volume shift introduced by these mutations could alter the interactions with the surrounding residues, and the FoldX analysis suggests a small ΔΔG(change) of -0.14 for these substitutions. The 135Pro>Ser mutation, a non-conservative substitution observed the Hg II sub-branch CCA2.9, is close to the protein-membrane interface ( Fig. 2 ). This mutation could potentially affect the mobility of the loop where it is located, promoting enhanced interactions with other subunits. FoldX analysis suggests that this mutation could be slightly destabilizing for the structure, with a ΔΔG (change) of about +0.99. COX3 - Cytochrome C Oxidase COX3 is found in the inner mitochondrial membrane and is one of the three mtDNA-encoded subunits of the respiratory complex IV. It is the terminal enzyme of the mitochondrial respiratory chain and catalyses the transfer of electrons from reduced cytochrome c to molecular oxygen. A unique mutation, 142Val>Met, was identified in the mt-COX3 subunit of cytochrome oxidase (Fig. 3A, B), which defines one of the Hg II sub-branches. This residue is in the protein core; its side chain protrudes at the interface with the neighbouring protein subunit and is involved in intra-subunit contacts. The valine-to-methionine substitution is conservative, preserving the hydrophobic character of the amino acid, but lengthens the side chain of the residue, allowing it to make additional/diversified contacts with surrounding subunits. FoldX analysis suggests this mutation could slightly stabilise the structure with a ΔΔG (change) of about -0.94 (Table 3B). CYTB - Cytochrome B CYTB is located in the inner mitochondrial membrane and is a component of the respiratory chain complex III. Five mutations have been recognized in the mt-CYTB subunit ( Fig. 4A ) four of which differentiate Hg IB from Hg II. All mutations are located within the phospholipid bilayer, in the protein transmembrane domain, on the helices forming the protein core. The 53Ala>Thr (sub-branch in Hg II) and Thr47Ile mutations are hydrophobic changes that involve a threonine residue into the membrane-embedded protein core ( Fig. 4B, C ). As expected, the analysis of the mutation energy shows that these mutations slightly stabilize the protein’s three-dimensional structure, with negative ΔΔG(change) of -2.73 when comparing Hg IB to Hg II basal ( Table 3A ) and -1.62 when comparing Hg IB to Hg II derived (Table 3B). The Val190Ile mutant represents a conservative variant that maintains a non-polar nature within the transmembrane structure but alters the residue volume. Similarly, the Phe182Leu and Phe233Leu mutations substitute an aromatic residue to the protein-membrane interface ( Fig. 4C ). Nuclear Overhauser Effect (NOE) spectroscopy studies have shown that leucine-to-phenylalanine mutations lead to much stronger interactions with non-polar interfaces (Mishra et al. 2008). In this hypothesis, the phenylalanine residues may increase the interaction between this subunit and the acyl chains of the membrane phospholipids, increasing the anchoring of the protein to the membrane in Hg IB. NADH dehydrogenase NADH dehydrogenase is an enzyme involved in the conversion of reduced nicotinamide adenine dinucleotide (NADH) to its oxidized form (NAD + ). Observations on this protein partially overlap with previous reports concerning homeoviscous adaptation (Novelletto et al. 2016). Notably, the ND1 subunit exhibited previously identified variants ( Thr73Ile, Ala314Thr, and Ser317Gly ) located at the transmembrane helix termini ( Fig. 5B ). Notably, isoleucine-to-threonine mutations located at the edges of transmembrane helices were shown to structurally stabilize transmembrane proteins and increase their thermostability (Zhou and Bowie 2000). These sites are also known to be involved in membrane thickness modulation (Novelletto et al. 2016). Moreover, a conservative Ile16Val variant located within the membrane was also observed. No significant perturbation of protein structure has been predicted by FoldX analysis of these mutants (Table 3A). Four mutations were observed in the ND2 subunit ( Fig. 5 C ), one of which within Hg II. The His4Tyr and Asn85Asp mutations are located at the membrane’s borders. In the first case, the mutation is a histidine-to-tyrosine substitution that maintains the presence of a comparable ring with “belt function”, for the amphipathic anchoring into the membrane, contributing to the structural stability of the protein in the bilayer. The second mutation, on the other hand, leads to substituting a polar residue with a charged one. Here, the asparagine amino group may form new hydrogen bonds with the polar heads of phospholipids, stabilizing the interaction with the membrane. Finally, the Ile31Thr mutation causes a small hydrophilic shift in the transmembrane portion. For these three mutations, FoldX predicts a slight destabilization of the protein (Table 3A). The 103Ala>Thr mutation is in the protein core and introduces a hydrophilic residue into a very apolar structural environment, which should slightly destabilize the protein structure, as also suggested by the FoldX ΔΔG (change) of +1.26 (Table 3B). For the ND3 subunit, two previously reported variants, Met5Thr and Val108Ile were identified when comparing Hg IB and Hg II, in addition to the novel 27Leu>Ser variant ( Fig. 5D ). The Met5Thr and 27Leu>Ser substitutions represent potential hydrophilic/hydrophobic (and vice versa ) changes that could affect the membrane thickness regulation. Conversely, the Val108Ile substitution appears to be conservative, only causing a small decrease in volume and likely preserving the protein functional role, due to its location within a highly conserved region. Nevertheless, also in this case the FoldX analysis predicts negligible effects on the protein resulting from these substitutions. A single mutation, Ser35Ala, was found on the ND4 subunit, resulting in a FoldX ΔΔG(change) of about -0.84 when comparing Hg IB to Hg II (Table 3A). This hydrophilic switch is mapped to the protein α-helices core within the transmembrane portion. No substitution was observed in this subunit within Hg II. The ND5 subunit showed the highest number of mutations, i.e. 8 in the Hg IB-Hg II basal comparison, increasing to 9 in the Hg IB-Hg II derived comparison. All of these are predicted as slightly destabilising by FoldX, with a ΔΔG(change) between 0.63 in the first comparison, increasing to 0.68 in the second one. Two hydrophobic shifts, Thr50Ala and 487Ala>Thr, and one hydrophilic shift, Leu572Ser, were found. These shifts are mapped to the protein’s boundary and are located close to the membrane’s edge. A Leu18Phe mutation can be related to the previous description of increased protein stability upon contact with non-polar surfaces. Finally, conservative mutations, Ala87Val and Val177Ile, were observed in the protein’s transmembrane portion, as well as mutations that substitute residues with a sulfur-containing group, Cys24Tyr, Met31Leu, and Met446Thr .

AND CONCLUSIONS The phylogenetic analysis of 61 complete sequences in our dataset ( Fig. 1 ) corroborates the previously established topology for the species. Our findings confirm that Hg IB is the most basal, with Hg IA forming a sister clade to Hg II. This relationship underlines the evolutionary branching within these haplogroups and reinforces the accuracy of our dataset. Fifty variants were identified in protein-coding regions among Hg II specimens, with 33 of them basal. The ubiquitous presence of these variants among Hg II carriers signifies their fundamental role in the haplogroup genetic makeup. The 9440 A/T variant, though not present in all Hg II carriers, remains a notable marker within this group. The detection of heteroplasmy involving the ND2 234Trp>Stop mutation is particularly noteworthy. This mutation represents a truncating variant, which is unlikely to be transmitted to offspring due to its severity and the strong evolutionary conservation of the corresponding position in the ND2 protein. This observation strongly suggests a selective disadvantage and highlights the stringent evolutionary constraints on mitochondrial DNA variants that significantly alter protein function. Molecular modelling has been performed to understand the significance of the observed residue substitutions. Analysis of the observed variants in the proteins within their 3D context did not identify critical structural problems, but suggested possible alterations caused by the substitutions. For example, the 135Pro>Ser, Thr78Ala, and Phe151Leu mutations within the ATP6 subunit (Fig. 2) would indicate potential functional impacts on oxidative phosphorylation, due to altered protein-membrane interactions. The 142Val>Met mutation in COX3 ( Fig. 3 ), while conservative, suggests slight modifications in protein-protein interactions within the cytochrome oxidase complex. Mutations in the CYTB subunit ( Fig. 4 ), particularly the hydrophilic shifts 53Ala>Thr and Thr47Ile, hint at minor changes to the protein’s structure, potentially affecting its location within the phospholipid bilayer. Overall, results derived from FoldX analysis between Hg IB and Hg II basal proteins indicate that the structures’ functionality are fundamentally preserved, as the calculated change in free energy fluctuates around zero and is repeatedly minimal ( Table 3 A). This finding is consistent with the long-term evolutionary viability of these substitutions in the two lineages whose divergence was estimated at 3-4 million years ago (Duchene et al. 2012; Clusa et al. 2013; Tolve et al. 2024). It is likely that molecular mechanisms which preserve mitochondria functionality are at work despite the appearance of possibly damaging alterations in protein structures. From one side, the mutation location and/or type may allow the protein structures to buffer the effect, thanks to the remarkable structural adaptability of membrane proteins, enabling them to cope with mutations without suffering significant functional consequences. For example, the various mutations observed in NADH dehydrogenase (ND1, ND2, ND3, ND4, and ND5 subunits) highlight adaptive changes, possibly related to membrane thickness, composition, and regulation. Moreover, isoleucine-to-threonine mutations, as found in ND1, located within transmembrane helices and close to the membrane edges, were shown to structurally stabilize and increase the thermostability of membrane proteins (Zhou and Bowie, 2000). Indeed, in many cases, except for the two mutations in the ND3 subunit, the variants show small positive ΔΔGs, indicating that the mutation only slightly destabilized the structure in Hg IB vs Hg II ( Table 3A ). This scenario suggests the presence of a homeostatic system capable of buffering possibly damaging structural changes, probably through membrane lipid remodelling (Breton et al. 2021). Indeed, mitochondrial membranes are dynamic and constantly remodel their structure and composition to maintain mitochondrial function and homeostasis. Structural alterations and change in composition of mitochondrial membranes, in response to mutations in membrane proteins, can affect membrane fluidity, permeability, and protein-lipid interactions, thus potentially compensating for the altered protein function by modifying the local environment around them. Further remodelling responses triggered by stress can involve changes in membrane protein expression, turnover rates, or recruitment of auxiliary proteins, which may help alleviate the impact of mutations by enhancing or replacing malfunctioning proteins. Finally, mitochondrial membranes may contain functionally redundant proteins or alternative pathways, which may be modulated during stress conditions and that can partially compensate for damaging mutations in essential membrane proteins. Another feature of mitochondria that may help in buffering potentially destabilizing mutations is heteroplasmy, i.e. the co-existence of different genomic variants in the same organelle simultaneously (Breton et al. 2021; Delgado-Cano et al. 2021). It seems that the extent of phenotypic expression of a particular variant depends on the abundance of the corresponding allele in the mitochondrial population, such that a certain mtDNA variant will effectively affect the phenotype only if it exceeds a certain frequency in the mitochondrial population. Therefore, more common variants may buffer the effects of negative variants (Breton et al. 2021). Indeed, a moderately high level of heteroplasmy has been detected in the nd4 and nd5 genes for C. caretta (Delgado-Cano et al. 2020; Delgado-Cano et al. 2021), the latter showing the highest number of mutations in our analysis ( Table 3 ). Mitochondrial fission and fusion, intracellular transfers, and the partition of mtDNA into daughter cells during cell division can influence the frequency of particular variants in the mitochondrial population, further altering the equilibrium between the appearance of different phenotypes. It has been shown that this mechanism may not be completely random and may lead to the selection or deletion of specific variants, which may favour or reduce species adaptation (Breton et al. 2021). Finally, other studies suggest that turtles can adapt to anoxic and cold conditions through a reduction in mitochondrial respiration rate and substrate oxidation, involving posttranslational modifications and changes in subunit composition (Bundgaard et al. 2019). Furthermore, an increase in the flexibility of mitochondrial membrane proteins due to mutations (slight increases in ΔΔGs following mutations) can be hypothesized as a mechanism to counteract the adverse conditions of elevated pressure experienced during deep-sea diving. At high depths (loggerhead turtles can dive up to -200 meters deep), the high pressure can destabilize cellular structures and impair protein function. However, mutations that enhance the flexibility of these proteins allow them to maintain their structural integrity and functional activity despite the intense pressure. This adaptability ensures that essential mitochondrial functions, such as ATP production and metabolic regulation, continue efficiently, thereby supporting cellular and organismal survival in extreme underwater environments. The 5 comparisons involving new variants, derived within Hg II, were all associated with ΔΔGs increased toward positive values (compare Tables 3A and 3B gene-by-gene). The negative ΔΔG value for COX3 could not be contrasted with a corresponding value in the Hg IB vs Hg II basal comparison, since the single variant was found within Hg II. These observations suggest that the novel mutations generally result in a slight destabilization of the protein structure within the mitochondrial membrane 3D context. Quantitatively the shifts of ΔΔGs rule out drastic effects on the proteins’ functionality as well as a generalized change in the pattern of stability as compared to the Hg IB vs Hg II long term divergence. The reason behind this trend may be the structural adaptability of proteins, as well as other mitochondrial remodelling mechanisms. In this context, some of the new variants might simply add to the Hg II basal ones in conferring a better performance in the Mediterranean conditions to the respective protein product. On top of this, it is conceivable that only a subset of the polymorphic variants are adaptive, reducing the power of the present analysis to capture generalized effects. At any rate, in view of the variety of the remodeling effects of all substitutions discussed above, homeoviscous adaptation cannot be called as the sole mechanism underlying potential adaptation to different climatic conditions. This was initially put forward based on an apparent excess of substitutions close to the boundaries of the lipid bilayer in ND1 and ND3 (Novelletto et al. 2016). Concluding, understanding the mechanisms behind the unique genetic adaptations of the loggerhead turtle C. caretta to diverse climatic conditions is critical from a conservation perspective, since disruptions in mitochondrial efficiency can significantly impact health, reproductive success, and survival rates. The present analysis provides a structural and energetic view of the evolutionary dynamics of protein-coding mitochondrial DNA and the potential functional implications of specific mutations in critical mitochondrial proteins, associated with the colonization of the Mediterranean and radiation within it. It adds to the understanding of mitochondrial evolution and the genetic diversity within and between C. caretta haplogroups. Complete mitogenome screenings of larger samples from the Mediterranean basin will help in revealing a larger array of uncommon variants, but also to assess their frequencies more precisely, enabling to evaluate their intra-specific dynamics and associated adaptive value.

Baltazar-Soares M, Klein JD, Correia SM, Reischig T, Taxonera A, Roque SM, Dos Passos L, Durão J, Lomba JP, Dinis H, Cameron SJK, Stiebens VA, Eizaguirre C (2020) Distribution of genetic diversity reveals colonization patterns and philopatry of the loggerhead sea turtles across geographic scales. Sci Rep 22: 18001 Bolger AM, Lohse M, Usadel B (2014) Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics 30: 2114-2120 Bowen B, Avise JC, Richardson JI, Meylan AB, Margaritoulis D, Hopkins-Murphy SR (1993) Population structure of loggerhead turtles ( Caretta caretta ) in the Northwestern Atlantic Ocean and Mediterranean Sea. Conserv Biol 7: 834-844 Bowen BB, Meylan AB, Perran Ross J, Limpus CJ, Balasz GH, Avise JC (1992) Global population structure and natural history of the green turtle ( Chelonia mydas ) in terms of matriarchal phylogeny. Evolution 46: 865-881 Breton S, Ghiselli F, Milani L (2021) Mitochondrial short-term plastic responses and long-term evolutionary dynamics in animal species. Genome Biol Evol 13 Bundgaard A, James AM, Gruszczyk AV, Martin J, Murphy MP, Fago A (2019) Metabolic adaptations during extreme anoxia in the turtle heart and their implications for ischemia-reperfusion injury. Sci Rep 9: 2850 Carreras C, Pascual M, Tomás J, Marco A, Hochscheid S, Castillo JJ, Gozalbes P, Parga M, Piovano S, Cardona L (2018) Sporadic nesting reveals long distance colonisation in the philopatric loggerhead sea turtle (Caretta caretta). Scientific Reports 8: 1435 Casale P, Margaritoulis D (2010) Sea turtles in the Mediterranean: distribution, threats and conservation priorities. IUCN, Gland, Switzerland, pp 294 Cingolani P, Platts A, Wang LL, Coon M, Nguyen T, Wang L, Land SJ, Lu X, Ruden DM (2012) A program for annotating and predicting the effects of single nucleotide polymorphisms, SnpEff. Fly 6: 80-92 Clusa M, Carreras C, Pascual M, Demetropoulos A, Margaritoulis D, Rees AF, Hamza AA, Khalil M, Aureggi M, Levy Y, Turkozan O, Marco A, Aguilar A, Cardona L (2013) Mitochondrial DNA reveals Pleistocenic colonisation of the Mediterranean by loggerhead turtles ( Caretta caretta ). J Exp Mar Biol Ecol 439: 15-24 Delgado-Cano D, Mariño-Ramirez L, Hernández-Fernández J (2020) Data on heteroplasmic mutations in mitochondrial genomes of loggerhead and hawksbill sea turtles: First approach. Data Brief 28: 104882 Delgado-Cano D, Mariño-Ramírez L, Hernández-Fernández J (2021) Detection of high heteroplasmy in complete loggerhead and hawksbill sea turtles mitochondrial genomes using RNAseq. Mitochondrial DNA A DNA Mapp Seq Anal 32: 106-114 Di Marino D, Oteri F, Morozzo Della Rocca B, Chillemi G, Falconi M (2010) ADP/ATP mitochondrial carrier MD simulations to shed light on the structural-dynamical events that, after an additional mutation, restore the function in a pathological single mutant. J Struct Biol 172: 225-232 Drosopoulou E, Tsiamis G, Mavropoulou M, Vittas S, Katselidis KA, Schofield G, Palaiologou D, Sartsidis T, Bourtzis K, Pantis J, Scouras Z (2012) The complete mitochondrial genome of the loggerhead turtle Caretta caretta (Testudines: Cheloniidae): Genome description and phylogenetic considerations. Mitochondrial DNA 23: 1-12 Duchene S, Frey A, Alfaro-Nùñez A, Dutton PH, Thomas P. Gilbert M, Morin PA (2012) Marine turtle mitogenome phylogenetics and evolution. Mol Phyl Evol 65: 241-250 Encalada SE, Bjorndal KA, Bolten AB, Zurita JC, Schroeder B, Possardt E, Sears CJ, Bowen BW (1998) Population structure of loggerhead turtle ( Caretta caretta ) nesting colonies in the Atlantic and Mediterranean as inferred from mitochondrial DNA control region sequences. Mar Biol 130: 567-575 Ernst R, Ejsing CS, Antonny B (2016) Homeoviscous adaptation and the regulation of membrane lipids. J Mol Biol 428: 4776-4791 Falconi M, Venerini F, Desideri A (1998) A spectroscopic and molecular dynamics study of native and of a mutant of Xenopus laevis Cu,Zn superoxide dismutase: mechanistic consequences of replacing four charged aminoacids on the “electrostatic” loop. Biophys Chem 75: 235-248 FitzSimmons NN, Limpus CJ, Norman JA, Goldizen AR, Miller JD, Moritz C (1997) Philopatry of male marine turtles inferred from mitochondrial DNA markers. Proc Natl Acad Sci USA 94: 8912-8917 Garofalo L, Mingozzi T, Micò A, Novelletto A (2009) Loggerhead turtle ( Caretta caretta ) matrilines in the Mediterranean: further evidence of genetic diversity and connectivity. Mar Biol 156: 2085-2095 Hawkes LA, Broderick AC, Godfrey MH, Godley BJ (2007) Investigating the potential impacts of climate change on a marine turtle population. Glob Chang Biol 13: 1-10 Hazel JR (1995) Thermal adaptation in biological membranes: is homeoviscous adaptation the explanation? Annu Rev Physiol 57: 19-42 Hernández-Fernández J, Delgado Cano DA (2018) Complete mitochondrial genome of the three Colombian Caribbean loggerhead sea turtles, using next-generation sequencing. Mitochondrial DNA Part B 3: 626-627 Langmead B, Salzberg SL (2012) Fast gapped-read alignment with Bowtie 2. Nat Methods 9: 357-359 Li H, Handsaker B, Wysoker A, Fennell T, Ruan J, Homer N, Marth G, Abecasis G, Durbin R (2009) The Sequence Alignment/Map format and SAMtools. Bioinformatics 25: 2078-2079 Loftus SJ, Crawford DL (2013) Interindividual variation in Complex I activity in Fundulus heteroclitus along a steep thermocline. Physiol Biochem Zool 86: 82-91 Márquez MR (1990) Sea turtles of the world. An annotated and illustrated catalogue of sea turtle species known to date. FAO Fisheries Synopsis No. 125, Rome Mazaris AD, Kornarali E, Matsinos YG, Margaritoulis D (2004) Modeling the effect of sea surface temperature on sea turtle nesting activities by investigating seasonal trends. Nat Res Model 17: 445-465 Mishra VK, Palgunachari MN, Krishna R, Glushka J, Segrest JP, Anantharamaiah GM (2008) Effect of leucine to phenylalanine substitution on the nonpolar face of a class A amphipathic helical peptide on its interaction with lipid: high resolution solution NMR studies of 4F-dimyristoylphosphatidylcholine discoidal complex. J Biol Chem 283: 34393-34402 Montaña-Lozano P, Balaguera-Reina SA, Prada-Quiroga CF (2023) Comparative analysis of codon usage of mitochondrial genomes provides evolutionary insights into reptiles. Gene 851: 146999 Napoli I, Mercaldo V, Boyl PP, Eleuteri B, Zalfa F, De Rubeis S, Di Marino D, Mohr E, Massimi M, Falconi M, Witke W, Costa-Mattioli M, Sonenberg N, Achsel T, Bagni C (2008) The fragile X syndrome protein represses activity-dependent translation through CYFIP1, a new neuronal 4E-BP. Cell 134: 1042-1054 Novelletto A, Testa L, Iacovelli F, Blasi P, Garofalo L, Mingozzi T, Falconi M (2016) Polymorphism in mitochondrial coding regions of Mediterranean loggerhead turtles: evolutionary relevance and structural effects. Physiol Biochem Zool 89: 473-486 Papadopoulos JS, Agarwala R (2007) COBALT: constraint-based alignment tool for multiple protein sequences. Bioinformatics 23: 1073-1079 Schymkowitz J, Borg J, Stricher F, Nys R, Rousseau F, Serrano L (2005) The FoldX web server: an online force field. Nucleic Acids Res 33: W382-388 Shamblin B, Bagley D, Ehrhart L, Desjardin N, Martin RE, Hart K, Naro-Maciel E, Rusenko K, Stiner J, Sobel D, Johnson C, Wilmers T, Wright L, Nairn C (2015) Genetic structure of Florida green turtle rookeries as indicated by mitochondrial DNA control region sequences. Cons Genet 16: 673-685 Shamblin BM, Bolten AB, Abreu-Grobois FA, Bjorndal KA, Cardona L, Carreras C, Clusa M, Monzòn-Arguello C, Nairn CJ, Nielsen JT, Nel R, Soares LS, Stewart KR, Vilaça ST, Turkozan O, Yilmaz C, Dutton PH (2014) Geographic patterns of genetic variation in a broadly distributed marine vertebrate: New insights into loggerhead turtle stock structure from expanded mitochondrial DNA sequences. PLoS ONE 9: e85956 Shen MY, Sali A (2006) Statistical potential for assessment and prediction of protein structures. Protein Sci 15: 2507-2524 Somero GN (2010) The physiology of climate change: how potentials for acclimatization and genetic adaptation will determine ‘winners’ and ‘losers’. J Exp Biol 213: 912-920 Stiebens VA, Merino SE, Roder C, Chain FJJ, Lee PLM, Eizaguirre C (2013) Living on the edge: how philopatry maintains adaptive potential. Proc Roy Soc B: Biological Sciences 280: 20130305 Strobel A, Graeve M, Poertner HO, Mark FC (2013) Mitochondrial acclimation capacities to ocean warming and acidification are limited in the Antarctic nototheniid fish, Notothenia rossii and Lepidonotothen squamifrons . PLoS ONE 8: e68865 Tolve L, Iannucci A, Garofalo L, Ninni A, Capobianco Dondona A, Cocumelli C, De Lucia A, Falconi M, Formia A, Iacovelli F, Mancusi C, Marchiori E, Marsili L, Mingozzi T, Nannarelli S, Natali C, Terracciano G, Zuffi MAL, Novelletto A, Ciofi C (2024) Whole mitochondrial genome sequencing provides new insights into the phylogeography of the Mediterranean loggerhead sea turtle ( Caretta caretta ). Mar Biol 171: 19 Waterhouse A, Bertoni M, Bienert S, Studer G, Tauriello G, Gumienny R, Heer FT, de Beer TAP, Rempfer C, Bordoli L, Lepore R, Schwede T (2018) SWISS-MODEL: homology modelling of protein structures and complexes. Nucleic Acids Res 46: W296-w303 Wilson AB, Eigenmann Veraguth I (2010) The impact of Pleistocene glaciation across the range of a widespread European coastal species. Mol Ecol 19: 4535-4553 Zhou Y, Bowie JU (2000) Building a thermostable membrane protein. J Biol Chem 275: 6975-6979 Zhu J, Vinothkumar KR, Hirst J (2016) Structure of mammalian respiratory complex I. Nature 536: 354-358

We are grateful to the following collaborators involved in sample collection: Nicola Novarini of the Natural History Museum of Venice, Lisa Poppi, Caterina Raso, Luca Marini and collaborators of the NGO TartaLazio and all voluntary workers and former students involved in the TartaCare Project of Calabria University including Sandra Berlingeri, Giulia Cambiè, Francesca Crispino, Concetta Denaro, Simona Fabiano, Elena Giardinazzo, Teresa Malito, Carmela Mancuso, Giampiero Masciari, Alessandro Massolo, Nunzia Micò, Gianni Parise, Brunella Pisani, Manuela Policastrese, Patrizia Rima, Manuela Russo, Salvatore Salice and Salvatore Urso. We also thank Luana Papetti and collaborators of the NGO tartAmare and Isa Tonso and collaborators of Legambiente Arcipelago Toscano for support during collecting pipped eggs from Tuscany nesting sites. Author contributions ANo, MF and CCi designed and conceived the study; AN, FI and AR performed the structural analyses; LG, ACD, CCo, ADL, CM, EM, LM, TM, SN, GT, MALZ were involved in sample collection; CCi, AF, LM and CM secured laboratory work and consumables funding; LT, AI and CN prepared samples for sequencing in the wet lab; ANo, FI, AR, AN, MF and CCi wrote the paper. All authors revised and approved the final manuscript. None of them declared conflicts of interests. Data availability Mitochondrial DNA sequences presented in this study were deposited in the GenBank database (accession numbers: OR166582 - OR166647). Compliance with Ethical Standards Sampling standards were described elsewhere (Tolve et al. 2024). The study complies with Institutional, national and international ethics guidelines concerning the use of animals in research. None of the procedures used in the study met the criteria to define them as ”experiments” as defined in Article 2 of the EEC Directive 86/609/EEC regarding the protection of animals used for experimental and other scientific purposes. Funding This study was funded by the Tuscany Regional Authority NATura Network Toscana agreement in compliance with the Habitats Directive 92/43/EEC. Additional support was provided by the Italian Ministry of University and Research through the National Biodiversity Future Center, part of the National Recovery and Resilience Plan, Mission 4, Component 2, Investment 1.4, Project CN00000033. LT was granted a graduate research scholarship supported by the European Structural and Investment Funds through the National Operational Programme (PON) of the Italian Ministry of University and Research. Figure legends Figure 1. Bottom right: unrooted maximum parsimony tree obtained with 61 complete C. caretta mitochondrial DNA sequences, modified from (Tolve et al. 2024). Branch length is proportional to the number of nucleotide substitutions along the entire molecule. Haplogroup nomenclature is reported to the right, with Hg II shaded. Previous works (Duchene et al. 2012) show that the root is to be placed on the branch leading to Hg IB. Top left: subtree of sequences belonging to Hg II. Both panels: for Hg II and the branch basal to it, all substitutions in protein-coding genes are shown next to the corresponding branch in the polarised form and coloured: red = non-synonymous; purple = stop gain; green = synonymous. The position of the reference sequence FR694649.1 is shown. Figure 2. ATP6 3D context and mutations: ( A-B) ATP6 structure in the 3D context of ATP-synthase structure ( A ) and zooming ATP6 ( B ) with mutate residue highlighted. ( C ) Modelling of mutated ATP6. Figure 3. COX3 3D context and mutations: ( A-B ) COX3 structure in the 3D context of Cytochrome C oxidase structure ( A ) and zooming ( B ) with the variant residue highlighted. ( C ) Modelling of mutated COX3. Figure 4. CYTB 3D context and mutations: (A) CYTB structure in the 3D context of Complex III structure ( left panel ); (B,C) zooming with residues found in Hg II highlighted. Figure 5. NADH dehydrogenase 3D context and mutations: ( A ) ND mitochondrial subunit structures in the 3D context of Cytochrome B structure ( upper panel ) and zooming ( lower panel ); ( B-F ) Modelling of mitochondrial ND subunits: ND1 (B), ND2 (C), ND3 (D), ND4 (E) and ND5 (F) . Table 1. List of non-synonymous protein coding variants among C. caretta Haplogroup II complete mitochondrial sequences. | Gene | Position 1 | Ancestral nucleotide state | Derived nucleotide state | AA. position in the protein | Ancestral aa. | Derived aa. | Position in Hg II philogeny | Exceptions to conservation of the ancestral aa. in other Chelonidae | | ATP6 | 8398 | C | T | 135 | Pro | Ser | Derived, basal to Hg CCA2.9 | Pro conserved in all species | | CYTB | 14373 | G | A | 53 | Ala | Thr | Derived, 2 specimens | Ala conserved in all species | | COX3 | 9102 | G | A | 142 | Val | Met | Derived, private | Ile in L. olivacea | | ND2 | 4223 | A | G | 85 | Asn | Asp | Basal | Asp in C. mydas, N. depressus | | ND2 | 4277 | G | A | 103 | Ala | Thr | Derived, private | Ala conserved in all species | | ND3 | 9610 | T | C | 27 | Leu | Ser | Derived, private | Met in E. imbricata | | ND3 | 9853 | G | A | 108 | Val | Ile | Basal | Multiple aa. across species | | ND5 | 11880 | C | T | 18 | Leu | Phe | Basal | Ile in L. olivacea | | ND5 | 11899 | G | A | 24 | Cys | Tyr | Basal | Multiple aa. across species | | ND5 | 11919 | A | C | 31 | Met | Leu | Basal | Multiple aa. across species | | ND5 | 13287 | G | A | 487 | Ala | Thr | Derived, private | Thr in N. depressus | 1 Numbered according to Acc. n. FR694649.1 Table 2. List of templates used for homology modelling and associated DOPE score | Subunit | PDB Template | Method | Organism Template | Sequence Identity | GMQE | DOPE Score | | atp6 | 6zpo.a | EM | Bos taurus | 57,96 | 0,78 | -25754.2 | | cytb | 3h1i.C | X-ray | Gallus gallus | 78,1 | 0,87 | -48463.1 | | cox3 | 20cc.C | X-ray | Bos taurus | 83,15 | 0,94 | -34104.1 | | nd1 | 5ldw.H | EM | Bos taurus | 72 | 0,71 | -38132.1 | | nd2 | 5ldw.N | EM | Bos taurus | 48,98 | 0,74 | -42637 | | nd3 | 6g2j.A | EM | Mus musculus | 53,04 | 0,75 | -8585.8 | | nd4 | 6zkp.M | EM | Ovis aries | 63,44 | 0,77 | -68858.8 | | nd5 | 7v2d.l | EM | Sus scrofa | 61,02 | 0,85 | -82007.4 | Table 3. ΔΔG values calculated for each protein when comparing Hg IB to FR694649 (A) and Hg IB to Hg II including the derived variants (B). Residues which characterize Hg II sub-branches are greyed. A | Gene | Position | Nucleotide in Hg IB | Nucleotide in FR694649 | AA position | Residue in Hg IB | Residue in FR694649 | ΔΔG | | ATP6 | 8227 | G | A | 78 | Ala | Thr | -0.14 | | 8448 | A | C | 151 | Leu | Phe | || | CYTB | 14356 | C | T | 47 | Thr | Ile | -2.73 | | 14760 | T | C | 182 | Phe | Leu | || | 14784 | G | A | 190 | Val | Ile | || | 14913 | T | C | 233 | Phe | Leu | || | ND1 | 2841 | A | G | 16 | Ile | Val | -0.57 | | 3013 | C | T | 73 | Thr | Ile | || | 3735 | G | A | 314 | Ala | Thr | || | 3744 | A | G | 317 | Ser | Gly | || | ND2 | 3980 | C | T | 4 | His | Tyr | 0.76 | | 4062 | T | C | 31 | Ile | Thr | || | 4223 | A | G | 85 | Asn | Asp | || | ND3 | 9544 | T | C | 5 | Met | Thr | 0.32 | | 9853 | G | A | 108 | Val | Ile | || | ND4 | 10343 | T | G | 35 | Ser | Ala | -0.84 | | ND5 | 11880 | C | T | 18 | Leu | Phe | 0.63 | | 11899 | G | A | 24 | Cys | Tyr | || | 11919 | A | C | 31 | Met | Leu | || | 11976 | A | G | 50 | Thr | Ala | || | 12088 | C | T | 87 | Ala | Val | || | 12357 | G | A | 177 | Val | Ile | || | 13165 | T | C | 446 | Met | Thr | || | 13543 | T | C | 572 | Leu | Ser | B | Gene | Position | Nucleotide in Hg IB | Nucleotide in FR694649 including variants | AA position | Residue in Hg IB | Residue in FR694649 including variants | ΔΔG | | ATP6 | 8227 | G | A | 78 | Ala | Thr | 0.99 | | 8398 | C | T | 135 | Pro | Ser | || | 8448 | A | C | 151 | Leu | Phe | || | CYTB | 14356 | C | T | 47 | Thr | Ile | -1.67 | | 14373 | G | A | 53 | Ala | Thr | || | 14760 | T | C | 182 | Phe | Leu | || | 14784 | G | A | 190 | Val | Ile | || | 14913 | T | C | 233 | Phe | Leu | || | COX3 | 9102 | G | A | 142 | Val | Met | -0.94 | | ND2 | 3980 | C | T | 4 | His | Tyr | 1.26 | | 4062 | T | C | 31 | Ile | Thr | || | 4223 | A | G | 85 | Asn | Asp | || | 4277 | G | A | 103 | Ala | Thr | || | ND3 | 9544 | T | C | 5 | Met | Thr | 0.43 | | 9610 | T | C | 27 | Leu | Ser | || | 9853 | G | A | 108 | Val | Ile | || | ND5 | 11880 | C | T | 18 | Leu | Phe | 0.68 | | 11899 | G | A | 24 | Cys | Tyr | || | 11919 | A | C | 31 | Met | Leu | || | 11976 | A | G | 50 | Thr | Ala | || | 12088 | C | T | 87 | Ala | Val | || | 12357 | G | A | 177 | Val | Ile | || | 13165 | T | C | 446 | Met | Thr | || | 13287 | G | A | 487 | Ala | Thr | || | 13543 | T | C | 572 | Leu | Ser | Information & Authors Information Version history Copyright This work is licensed under a Non Exclusive No Reuse License.

Authors Metrics & Citations Metrics Article Usage 329views 265downloads Citations Download citation Andrea Ninni, Livia Tolve, Federico Iacovelli, et al. Structural Insights into Functional Impacts of Common Variants in Mitochondrial Protein-Coding Genes of Mediterranean Loggerhead Turtles. Authorea. 02 January 2025. DOI: https://doi.org/10.22541/au.173582686.60979887/v1 DOI: https://doi.org/10.22541/au.173582686.60979887/v1 If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download. For more information or tips please see 'Downloading to a citation manager' in the Help menu.